The major party leaders have made some curious statements regarding the formation of a government in the aftermath of an election where no party claims the majority of seats. First Thomas Mulcair, then Stephen Harper, and then Justin Trudeau as well, have asserted that whichever party wins more seats than the others should be called on by the Governor General to try and form a government. But that’s not what constitutional conventions, as understood by anyone who has actually studied them, say.
The conventions of responsible government require the government to enjoy the confidence of the House of Commons; but an incumbent a government is, we might say, deemed to be enjoying the confidence of the House until a vote of the House proves otherwise. The prorogation crisis of 2008, when the Governor General took the Prime Minister’s advice to prorogue Parliament is a case in point ― even though everybody knew that at that moment the government did not actually have the confidence of a majority of MPs, it is not enough for everybody to know. There has to be a Parliamentary vote to confirm what “everybody knows.” Similarly, it has always been said, after an election, the incumbent government is entitled to “meet the House” of Commons and test its confidence. Only if it does not obtain the confidence of the House does convention force it to resign.
Thus, contrary to what Messrs Mulcair, Harper, and Trudeau have claimed, the government’s number of seats has nothing to do with its entitlement to remain in office ― so long as no other party has got a majority. (If one has, there is no point in an incumbent government trying to meet the House, and I think it’s fair to say that there is a convention that demands the government’s prompt resignation in that event.) Accordingly, Andrew Coyne is wondering whether the leaders “understand the basics of our system of gov[ernmen]t,” and Emmett Macfarlane is positive that they do not. Dale Smith thinks they are “totally wrong.” Others have chimed in to the same effect as well. But the better question, in my view, has been asked by John Michael McGrath: “At what point does this go from ‘wrong answer’ to ‘new unwritten convention’?”
Constitutional conventions are just like law in that they are only valid until they have been changed. Conventions are unlike (some) law in that it is often difficult to tell when they have changed. (Actually, the common law, or any law articulated by courts is sometimes like that too. It is only at a certain point in the development of a line of cases that it becomes possible to say that the law is no longer what it had been at some earlier point.) But while it is can be difficult to pinpoint the moment when a convention has changed, we do know that conventions can change, and those pertaining to government formation are no exception.
Conventions, W.S. Holdsworth wrote in “The Conventions of the 18th Century Constitution,” (1932) 17 Iowa L. Rev. 161 ensure “that the constitution works in practice in accordance with the prevailing constitutional theory of the time.” (163) As the prevailing constitutional theory changes, so do conventions. How do we know though that a convention has changed? The same way that we know that one exists in the first place ― by looking at the behaviour of political actors to determine whether, in Sir Ivor Jennings’ classic formulation (endorsed by the Supreme Court in the Patriation Reference), there are “precedents” for an alleged convention; whether “the actors in the precedents believe that they were bound by a rule”; and whether there is “a reason for the rule.” If the convention observed at point B is not the same is that which regulated the same events at some previous point A, the convention has changed.
To come back to Mr. McGrath’s question, how do we know whether the convention regulating government formation in Canada has changed? Prof. Macfarlane has suggested that we would know “[w]hen [an] incumbent wants to test confidence but a [Governor General] ignores them & appoints a new gov[ernmen]t before the House meets. Maybe.” I don’t think that’s the only possibility. We don’t need to wait for a Governor General to dismiss an incumbent government when an opposition party wins a majority of seats to know that convention requires the incumbents to resign in such circumstances. Incumbents unfailingly resign on their own in such circumstances. Similarly, we wouldn’t need the Governor General to dismiss a government that wants to appoint one formed by the party with a plurality of seats.
If the Conservatives do not win a plurality of seats on October 19, and the government resigns without trying to meet the House, despite no opposition party having won a majority, and especially if it explains its resignation by reiterating Mr. Harper’s belief that it would be improper for a party that has not won a plurality to attempt to govern, that will be a very important indication that the convention has changed. We would have a precedent, and we would have a statement from the actor in the precedent that he felt bound by a rule. Would there be a reason for the rule? Arguably, yes, though that’s a somewhat trickier question. The rule that the winner of a plurality of seats gets first crack at forming a government has simplicity to recommend it, and it appeals to our majoritarian intuitions. It is similar to the rule we use in our electoral system ― though ironically the opposition parties might be looking to change that. I’m not saying, mind you, that this rule would be a better one than the old one. Only that there would be some reasons to justify it.
Now, it has to be possible to say, as Mr. Coyne, prof. Macfarlane, and others do, that the political actors are simply mistaken as to what the applicable rule is, just as it must, I think, be possible to say that the Supreme Court has wrongly decided a case (constitutional or otherwise). Mr. Trudeau, for instance, is just wrong to claim that “that’s the way it’s always been, whoever commands the most seats gets the first shot at governing.” It hasn’t always been that way. But just as with a Supreme Court decision, a mistake, at least if it is repeated in the future, can generate a rule just as surely as a correct statement of the relevant norm. To twist Justice Jackson’s well-known statement somewhat, the Supreme Court ― and, in the case of conventions, the political leaders of the day ― may not be final because they are infallible, but their being final does give them a form of deemed infallibility.
Constitutional rules change. Neither the process of change nor its results are always pretty. That is true of formal constitutional amendments and of the decisions of the Supreme Court that have the same effect. It is also true of the change in constitutional conventions. And of course there is nothing improper with trying to resist or even to undo a change one does not like. But it is important to recognize the possibility of constitutional change, and not only to criticize those engaged in it as mistaken. The Canadian constitution might be undergoing a metamorphosis before our eyes ― whether that word makes us think of Richard Strauss or of Franz Kafka.